To increase the number of replicas you can set the pool size according to your requirements:
ceph osd pool set size 4
The placement of the copies (chassis) is called failure domain. This is configured in the ruleset you're using. You can change the ruleset for a given pool:
# get current ruleset for given pool
ceph osd pool get iscsi-pool crush_rule
crush_rule: replicated_rule
# dump ruleset
ceph osd crush rule dump replicated_rule
The docs also describe how to change a crush rule and modify the crushmap. Changing the data placement will cause a remapping of the PGs, depending on your ceph version it will remap 5% of misplaced PGs at max. The remapping process can be controlled with these osd config settings:
osd_recovery_max_active
osd_max_backfills
Set them to higher values to increase the recovery speed, but set them back to the defaults after you're finished.